Abstract
Hypotension after non-cardiac surgery is common and associated with harm. Anaesthetists treat hypotension in the post-anaesthesia care unit (PACU) with intravenous (IV) fluids and vasopressor medications. Our aim was to determine the incidence of hypotension after these treatments. We conducted a single centre retrospective cohort study of all adult patients who were hypotensive (systolic blood pressure less than 90 mmHg) in the PACU after non-cardiac, non-obstetric surgery over a one-year period. The primary outcome was a composite of hypotension or vasopressor infusion in the 24 h after PACU discharge. During the study 459 patients were hypotensive in the PACU. No treatment was administered in 232 (51%) episodes, IV fluid alone was administered in 138 (30%) episodes, vasopressors alone were administered in 22 (5%) episodes, and both fluid and vasopressors were administered in 67 (14%) patients. A total of 167 patients (36%) met the primary outcome, of which 118 (25%) were hypotensive and 49 (11%) required vasopressor infusions. The treatment group was significantly associated with the primary outcome (P < 0.001), with 36 (15%) patients who received no treatment becoming hypotensive, compared with 67 (46%, P < 0.001) patients who received IV fluid alone, 12 (55%, P < 0.001) who received vasopressors alone and 52 (75%, P < 0.001) who received both IV fluid and vasopressors. Patients who were hypotensive in the PACU frequently developed later hypotension or required vasopressors in the 24 h after PACU discharge. Treatments delivered in the PACU had limited long-term effectiveness. Novel treatments to protect patients from subsequent hypotension are urgently needed.
Introduction
Hypotension after surgery is common, and is associated with an increased risk of myocardial injury, 1 stroke, 2 kidney injury, 3 delirium 4 and mortality. 5 Even brief exposures to hypotension are associated with poor outcomes, and postoperative hypotension is associated with harm regardless of how it is defined. 6 Hypotension is a common problem in the post-anaesthesia care unit (PACU), and anaesthetists are commonly required to manage this problem. 7 In some cases postoperative hypotension is expected and planned for, but in the majority of patients it is an unexpected event. 8 Known risk factors for postoperative hypotension include American Society of Anesthesiologists Physical Status (ASA) classification of 3 or more, 7 use of beta blockers, intraoperative hypotension, longer duration of surgery, 9 the use of postoperative neuraxial analgesia 10 and completion of surgery after-hours. 11 Despite knowledge of these risk factors, prediction of postoperative hypotension remains difficult. 12 There are no validated prediction tools for postoperative hypotension in the PACU nor on the surgical ward, and experts advise that all surgical patients should be considered at risk. 13
As large-scale randomised controlled trials are lacking, anaesthetists must use clinical judgement about treatment and discharge to the surgical ward or intensive care unit (ICU). 14 Intravenous (IV) fluids and vasopressors are commonly employed to correct hypotension. The success of these treatments in preventing subsequent hypotension and need for emergency medical responses after PACU discharge have not been well characterised. There is also a paucity of literature describing the incidence of death within one year after an episode of hypotension in the PACU.
This study aimed to characterise the treatment choices made by anaesthetists when treating hypotension in the PACU, as well as examine the need for referral for postoperative ICU support. We also aimed to determine the incidence of subsequent hypotension and medical emergency team (MET) activation, and death within one year of PACU discharge, in adult patients after non-cardiac, non-obstetric surgery.
Methods
We performed a retrospective cohort study of all adult patients who were hypotensive (defined as any systolic blood pressure less than 90 mmHg recorded in the medical record) in the PACU after non-cardiac, non-obstetric surgery at the Royal Melbourne Hospital, a tertiary referral and trauma centre in Victoria, Australia, in the 2022 calendar year. The Melbourne Health Human Research Ethics Committee provided approval for the study (Approval Number: QA2022.133, 18/01/2023). Individual patient consent was waived. The study was performed in accordance with the Declaration of Helsinki and National Health and Medical Research Council code of practice.
We included patients aged 18 years and older who had at least one systolic blood pressure recorded in the PACU that was less than 90 mmHg. We excluded patients who had cardiac or obstetric procedures, who were transferred from the ICU to the operating theatre and were returning to the ICU after their surgery, those who were planned for direct ICU admission (for example following major aortic surgery) and those having non-surgical procedures such as gastrointestinal endoscopy, bronchoscopy, endovascular clot retrieval or bone marrow aspiration. Where patients underwent multiple procedures that were associated with hypotension in the PACU during the study period we included only the first procedure.
Patients were identified and data were manually extracted into an Excel spreadsheet (Microsoft Corporation, Redmond, WA, USA) from the hospital’s electronic medical record (EPIC Systems, Verona, WI, USA). Blood pressure was measured via either a non-invasive blood pressure cuff on the upper arm or an invasive radial arterial line in the PACU, which was removed prior to transfer to the ward, using a General Electric B450 physiological monitor (GE Healthcare, Mascot, NSW, Australia) by trained nurses in the PACU and the surgical wards. Outcome data were extracted by two investigators after training, which included discussion of variable coding and thresholds, example cases and feedback. Regular discussions were held during the extraction process to clarify points of difference. No formal testing of inter-rater reliability was undertaken.
We defined treatment of hypotension as any documented administration of IV fluid or vasoactive medications or treatment of any arrhythmia. We identified the treatments delivered by reviewing medical and nursing notes, the medication chart and records from any emergency responses triggered in the PACU. We recorded any IV fluids administered (including volume and type), vasopressors (by specific medication prescribed) or a combination of fluids and vasopressors.
The primary outcome was a composite of any hypotension (as defined above) or any initiation of a vasopressor infusion in the 24 h after PACU discharge. Secondary outcomes included unplanned referral to the ICU from the PACU, the destination after a MET call (remaining on the ward, transfer to ICU, death) and the incidence of mortality within one year of surgery.
We studied a convenience sample of one calendar year of data, anticipating there would be several hundred hypotension events from available audit data.
The primary and secondary outcomes are presented descriptively. Comparisons of event rates between binary outcomes were made using Pearson’s chi-square test or Fisher’s exact test as appropriate. Statistical significance was identified by a P value less than 0.05, with a Bonferroni correction applied to minimise the risk of type I error when conducting multiple analyses. For this purpose, an initial threshold of significance of P < 0.05 was adjusted to P < 0.004 given 12 separate analyses were conducted. Data analysis was undertaken using Stata17 (Stata Corporation, College Station, TX, USA).
Results
The hospital performed approximately 16,000 procedures over the study period and there were 459 (2.9%) patients who had at least one episode of hypotension in the PACU. The median age was 55 (interquartile range (IQR) 35–70) years, 257 (56%) of the patients were female and 252 (55%) had an ASA classification of 3 or higher. The median procedure duration was 114 (IQR 51–215) min and 87 (3.9%) procedures were completed during an after-hours period.
No treatment was administered in 232 (51%) patients, fluid alone was administered to 138 (30%), vasopressors alone were administered to 22 (5%) and both fluid and vasopressor were administered to 67 (14%). Full details of provided treatments and outcomes are available in Table 1.
Treatments administered in the post-anaesthesia care unit, N = 459.
Patients often received multiple interventions which were unable to be isolated. Consequently, totals may add to more than 100% in this table.
ICU: intensive care unit; PACU: post-anaesthesia care unit; MET: medical emergency team; IQR: interquartile range; IV: intravenous
A total of 167 patients (36%) met the primary outcome, of whom 118 (25%) were hypotensive on the surgical ward and 49 (11%) required vasopressor infusions in the ICU. The treatment group had a higher incidence of hypotension on the ward (P < 0.001), with 36 (15%) patients who received no treatment becoming hypotensive compared with 67 (46%, P < 0.001) patients who received IV fluid alone, 12 (55%, P < 0.001) who received vasopressors alone and 52 (75%, P < 0.001) who received both IV fluid and vasopressors. The incidence of hypotension was not statistically significantly different between patients receiving IV fluid and vasopressors compared with vasopressors alone (52 (75%) versus 12 (55%), P = 0.095). Compared to receiving IV fluid, patients were not statistically significantly more likely to be hypotensive if they were provided vasopressors alone (67 (49%) vs 12 (55%), P = 0.56) but were significantly more likely to be hypotensive if they received both IV fluid and vasopressors (67 (49%) vs 52 (75%), P < 0.001). Full details of the distribution of the primary outcome between the groups are provided in Figure 1.
Flow of patients with respect to the primary outcome. PACU: post-anaesthesia care unit; ICU: intensive care unit.
A total of 77 patients (17%) were referred to the ICU from the PACU. Of these, 46 (60%) were accepted from the PACU and the remaining 31 (40%) were transferred to the ward, either because their condition improved or because ICU admission was inconsistent with the goals of care. ICU admission from the PACU did not occur in the no treatment group but did occur more frequently in the vasopressor alone group than in the IV fluid alone group (10 (45%) vs 1 (0.7%), P < 0.001), but not statistically significantly between the IV fluid and vasopressor group compared with the vasopressor alone group (10 (45%) vs 35 (51%), P = 0.624). Of the 31 patients referred to ICU who were transferred to the ward, six (19%) deteriorated to the point of requiring MET activation within 24 h of PACU discharge and four (13%) were admitted to ICU following that deterioration. A further 13 (2.8%) patients were not referred to ICU in the PACU and subsequently deteriorated on the ward with hypotension requiring MET activation, and of those, two (16%) required ICU admission. Two patients (0.4%) were transferred to the ward from the PACU without referral to ICU in accordance with their goals of care.
A total of 123 (26%) patients left the PACU with altered MET call criteria after their hypotensive episode. Of these, 45 (37%) had received no treatment, 64 (52%) had received IV fluid alone, seven (5.7%) had received vasopressors alone and seven (5.7%) had received IV fluid and vasopressors. Twelve (10%) of these patients had been referred to ICU, four (3.3%) experienced a MET call and one (0.8%) was subsequently admitted to ICU from the surgical ward.
At one year, 49 (10%) patients had died, of whom 18 (8%) did not receive treatment in the PACU for hypotension and 31 (14%) did receive treatment in the PACU for hypotension (P < 0.001). Post hoc testing did not demonstrate any statistically significant difference in the incidence of death between the active treatment groups (P = 0.286).
Discussion
This study examined the treatments deployed and the outcomes for patients treated for hypotension in the PACU. The results reveal a substantial burden of ward hypotension despite treatment and imply that the treatments deployed do not result in a sustained normalisation of blood pressure during the first 24 h after PACU discharge. These results indicate that the physiological disturbances of surgery and anaesthesia persist after discharge from the PACU, and that patients who are hypotensive in the PACU have a substantial risk of experiencing hypotension on the ward post discharge from the PACU. In addition, the similar incidence of hypotension after any active treatment suggests a pressing need to develop more effective treatments for hypotension in the PACU, with a particular focus on developing treatments that are more durable in effect and do not require ICU admission. We were unable to find publications with which to compare our work.
While greater treatment intensity, as demonstrated by larger doses of IV fluid or the need for vasopressor medications, appeared to be associated with a higher incidence of subsequent hypotension, these differences did not reach statistical significance. The most likely explanation is that patients with a greater degree of physiological disturbance required more intensive treatment before normalisation of blood pressure was achieved, reflected in the higher rates of ICU admission observed in the patients who received vasopressors. This finding may also reflect the limited duration of effect of all of the treatments employed, with patients being discharged from the PACU before the effect of the administered treatments wore off. This effect may also be due to the ‘titration paradox’, where the influence of treatments on the patient’s blood pressure appears to be inverted until the correct treatment dose is identified. 15 The finding that the incidence of death at one year was higher in the patients actively treated for hypotension is not surprising, and likely reflects more deranged physiology in those requiring active treatment, perhaps with concomitant cardiovascular risk factors and sequelae. There is also a risk implied by these data that treatment with fluid, vasopressors or both may cause harm and that the difference in outcomes observed between the groups was due to the interventions provided. The non-interventional nature of this study means that untangling this potential confounding issue was impossible, and the observation should prompt further randomised trials of interventions to address hypotension in the PACU.
Our study also reveals an unexpectedly low rate of MET activation and subsequent ICU admission when recurrent hypotension occurred. This may illustrate potential lower significance attached to hypotension in surgical patients, capacity constraints on the MET system or ICU in our hospital, or the fact that nearly one-quarter of patients who were hypotensive in the PACU had alterations to the goals of care placed to facilitate discharge from the PACU. This is an important clinical finding, suggesting that patients remain exposed to a large burden of potentially harmful hypotension without the routinely expected critical care support being implemented.
There are several limitations to this work. The most important is that as a single-centre study the results may not be generalisable to all patient populations or settings. The study also is not able to draw conclusions about the clinical impact of hypotension, nor the relative contribution of PACU hypotension to the development of ward hypotension owing to the absence of non-hypotensive comparator patients. We also did not risk-stratify patients. A further consideration is that for the non-significant results of the multiple secondary comparisons, the correction for multiple comparisons may have been excessively conservative and resulted in type II error. Further large-scale work with greater power to detect outcomes will be necessary to establish whether this is the case.
Conclusion
Patients who were hypotensive in the PACU had a high incidence of hypotension or vasopressor requirement in the 24 h after leaving the PACU. Treatments delivered in the PACU were only effective for a limited period of time and the MET call system was incompletely utilised in response to later hypotension on the ward. Studies of more durable interventions to treat hypotension in the PACU should be a research priority, in the hope of reducing the risk of subsequent hypotension and its associated risk of myocardial injury, stroke, kidney injury, delirium and death.
Footnotes
Author Contribution(s)
Declaration of conflicting interests
The authors have no conflicts of interest to declare.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
